FRBSF Economic Letter
1996-11 | March 29, 1996
This Letter summarizes the papers and discussion at a conference that was held at the Federal Reserve Bank of San Francisco on March 1 under the joint sponsorship of the Bank and Stanford University’s Center for Economic Policy Research.
The conference papers (listed at the end) fell into the broad categories of the measurement and the management of policy. In the first category, one paper presented a method to measure policy-induced changes in the money supply and the effects that these changes have on interest rates. Another looked at some of the problems associated with a popular technique used to measure shocks to policy and to estimate the effects of these shocks on the economy.
In the second category, two papers involved issues of economic management in an environment where monetary policy is committed to maintaining fixed exchange rates. A third paper examined the effects of a less rigid form of commitment, specifically, one where the monetary authority commits itself to a pre-announced rate of money growth over a relatively short time horizon.
Looking for the liquidity effect
James Hamilton’s paper takes up one of the key questions in macroeconomics: How do interest rates respond when the money supply changes? While it is usually assumed that a monetary expansion would lead to lower interest rates (the familiar “liquidity effect”), in theory the response could go either way: If the monetary expansion were accompanied by an increase in inflation expectations, for example, the result would be higher interest rates; but if price expectations did not react immediately, interest rates would fall. Hamilton examines this issue empirically by isolating so-called “exogenous” changes in the money supply, that is, changes in the money supply that are not the result of other macroeconomic developments. Isolating the response to an exogenous factor is important, since otherwise any change in interest rates we observe could represent the common response of the money supply and interest rates to some third factor, rather than the response of interest rates to monetary policy.
Hamilton uses a measure of unpredictable changes in the Treasury’s balances with the Fed to isolate exogenous changes in the supply of reserves (which are a component of the money supply that the Fed controls directly). When a check drawn upon a private bank is deposited in the Treasury’s account with the Fed, the result is a reduction in bank reserves. While the Fed tries to insulate the quantity of reserves from swings in the Treasury’s balances, it obviously cannot offset swings it cannot predict. Such swings should have a measurable impact on reserves, and Hamilton carries out some empirical tests to show that this is indeed the case.
He also shows that these fluctuations matter for the federal funds rate. According to his calculations, an unanticipated $1 billion increase in Treasury balances reduces nonborrowed reserves by $440 million and — provided it occurs when the Fed cannot offset it — causes the federal funds rate to go up by 10 basis points. Thus, Hamilton concludes that there is a liquidity effect in the federal funds market.
Another way of measuring monetary policy’s effect on the economy is with statistical models called Vector Autoregressions (VARs). In these models the federal funds rate is typically assumed to be the instrument of policy. Movements in the funds rate that cannot be explained by the other variables in the model, such as inflation and real output, are defined to be policy shocks. Economists have studied how different variables in the model react to these unpredicted changes in the funds rate in order to understand how the economy reacts to monetary policy.
Glenn Rudebusch takes issue with this approach, arguing that the estimated funds rate equations do not accurately describe how the Fed behaves. Among other things, he points out that studies that attempt to model Fed behavior by estimating “reaction functions” show that the Fed’s behavior is more complex than that suggested by VARs. For instance, in contrast to the equations estimated in the VARs, the estimated reaction functions tend to change over time, apparently reflecting changes in how the Fed reacts to developments in the economy. In addition, estimated VARs also appear to be based on both too little and too much information: They do not include important economic variables that the Fed reacts to, but they do include data that are the latest available at the time of estimation, whereas policy makers have to act on the basis of preliminary data.
Rudebusch argues that these models do not lead to good measures of policy shocks, either. To demonstrate this, he develops an alternative measure of these shocks, which is based on financial markets’ forecasts of the funds rate. The advantage of using such a measure is that financial market forecasts will incorporate all available information and also account for the way the Fed’s reactions change over time. Fortunately, the appropriate data are available directly from the funds rate futures market. A measure of policy innovations is then obtained by subtracting the federal funds futures rate for a given period from the actual funds rate over that period. Rudebusch finds that this measure and the VAR-based measure are essentially uncorrelated with each other, which suggests that the VAR-based measures are not good measures of policy shocks.
The first of the papers on monetary policy management is Guillermo Calvo’s discussion of the recent collapse of the Mexican peso. The usual explanation for this episode focuses on Mexico’s large and growing current account deficit before the collapse. According to this explanation, the deficits was growing too large to be sustained in the long run and, absent other adjustments, the peso had to fall to correct the imbalance. Calvo feels that the current account deficit was not the entire explanation; for instance, he points out that few people thought much of the deficits just a few months before the collapse. According to Calvo, developments in the financial markets, as well as the behavior of the monetary authorities, were crucial ingredients in the collapse.
The trouble began when U.S. rates began to rise in 1994. This caused Mexico’s foreign exchange reserves to fall just as the Mexican treasury increased the issuance of short-term, dollar-indexed debt (“tesobonos”), exposing Mexico to the possibility that it would not be able to meet all short-term claims if they were not rolled over or if additional financing were not forthcoming. Although the reserves could have been countered by raising domestic interest rates, such an action would have compounded the contractionary pressures arising from the increase in U.S. interest rates, especially by aggravating the problems in Mexico’s banking sector. The Mexican authorities chose instead to expand domestic credit.
This policy exacerbated the pressures on the peso, and Mexico had to devalue in December 1994. However, what was intended as a limited devaluation turned into a run on the peso, and the authorities were forced to let it float. According to Calvo, the initial devaluation raised doubts about Mexico’s ability to repay its debt and led to a speculative attack on the currency. This event also alerted investors to the possibility that the same thing could happen in other emerging market economies with fixed exchange rates, such as Argentina, causing speculative attacks against their currencies, a development Calvo labels the “tequila” effect.
In some ways, the quandary that Mexican policy makers faced during late 1994 is familiar: Commitment to a fixed exchange rate requires that policy makers give up the use of monetary policy to stabilize the economy. In other words, flexible exchange rates provide national governments with a tool to stabilize the economy; with a fixed exchange rate, they must find another instrument to achieve these goals. Indeed, one of the key issues facing policy makers contemplating the forthcoming European monetary unification is how to compensate for the resulting loss of a policy instrument.
Willem Buiter and Kenneth Kletzer take up this issue in their paper. They first show that with flexible exchange rates, policy makers can use monetary policy to stabilize individual economies in response to adverse supply shocks, such as a sharp increase in the price of oil. However, using monetary policy this way has some distributional implications in their model of the economy; an unexpected increase in the rate of inflation, for instance, transfers wealth from creditors to debtors, and hence affects the rate of capital accumulation. Lump sum transfers can then be used to offset these effects.
The situation gets more complicated in a fixed exchange rate regime. Note first that countries under consideration now effectively have a common money supply. Buiter and Kletzer show that when an adverse shock hits, variations in the common money supply and lump sum transfers can be used to stabilize the economy. However, these policies do not make up for the loss of independent monetary policies in the individual countries, since they do not allow policy makers to attain the same distributional and capital accumulation goals as before. To attain these goals, policy makers have to employ distortionary fiscal policies — such as a payroll tax. Thus, Buiter and Kletzer show that it is possible to use fiscal policy instruments to make up for the loss of monetary policy; however, these instruments are likely to be more complicated than the simple lump sum transfers that some others have suggested in the context of the Maastricht treaty.
From one perspective, fixed exchange rates are a way to commit monetary policy to a certain course of action. V.V. Chari, Lawrence Christiano, and Martin Eichenbaum ask a more basic question about commitment: What happens if policy makers have complete discretion? Their answer is that in that case monetary policy itself can become a source of instability in the economy. A simple example illustrates the point. Suppose that economic agents (such as households and firms) erroneously come to believe that monetary policy will become expansionary in the future. They will react by raising prices and wages in anticipation. Policy makers are then faced with the following dilemma: If they refuse to validate these expectations and maintain an unchanged policy, the economy will go into a recession, because firms will not be able to sell all they want at the higher prices and workers will not be able to find employment at the higher wages; alternatively, if they choose to validate these expectations by adopting a more expansionary policy, the result will be higher inflation. It is not hard to imagine a benevolent policy maker responding to this tradeoff by choosing to accept higher inflation instead of forcing the economy into a recession. Thus, inflation can go up simply because agents expect it to. Similar problems also can arise if agents form erroneous expectations about how policy makers will react to fundamental shocks to the economy, such as the oil price increases of the 1970s.
One obvious solution is to have the monetary policy makers commit to a sequence of policy actions into the indefinite future, thereby eliminating all uncertainty about their future behavior. For instance, the monetary authority could commit to a simple rule, such as maintaining annual money growth at 5 percent forever. However, if the velocity of money shifted, a 5 percent rate of growth of money could be associated with economic performance that was very different from what policy makers had intended. To avoid this kind of problem, policy makers would have to spell out how they would behave under every possible set of circumstances into the indefinite future. This process would be difficult, expensive, and perhaps impossible. In light of these considerations, the authors propose that policy makers commit to a course of action for the near future. The commitment horizon would depend upon the planning horizon of agents in the economy. For instance, in an economy where 3-year wage contracts were the norm, policy makers would pre-announce money growth for the next 3 years; doing so would eliminate uncertainty about policy from the wage-contracting process.
These papers will be published as Working Papers.
Buiter, Willem H., and Kenneth M. Kletzer. “Monetary Union and Macroeconomic Stabilization.” UC Santa Cruz.
Calvo, Guillermo A. “Capital Flows and Macroeconomic Management: Tequila Lessons.” University of Maryland.
Chari, V.V., Lawrence J. Christiano and Martin Eichenbaum. “Expectational Traps and Discretion.” University of Minnesota.
Hamilton, James D. “Measuring the Liquidity Effect.” UC San Diego.
Rudebusch, Glenn D. “Do Measures of Monetary Policy in a VAR Make Sense?” Federal Reserve Bank of San Francisco.
Opinions expressed in FRBSF Economic Letter do not necessarily reflect the views of the management of the Federal Reserve Bank of San Francisco or of the Board of Governors of the Federal Reserve System. This publication is edited by Sam Zuckerman and Anita Todd. Permission to reprint must be obtained in writing.
Please send editorial comments and requests for reprint permission to
Attn: Research publications, MS 1140
Federal Reserve Bank of San Francisco
P.O. Box 7702
San Francisco, CA 94120